Systematic evaluation of the Precision ID GlobalFiler™ NGS STR panel v2 using single-source samples of various quantity and quality and mixed DNA samples.

Massively parallel sequencing (MPS) techniques were developed approximately 15 years ago. Meanwhile, several MPS kits for forensic identification, phenotypic information, ancestry, and mitochondrial DNA analysis have been developed and their use has been established. Sequencing short tandem repeats (STRs) has certain advantages over the currently used length-based genotyping methods, which are based on PCR amplification followed by capillary electrophoresis (CE). MPS is more discriminative and includes the possibility of testing high numbers of targets (> 100), different types of markers [STRs and single nucleotide polymorphisms (SNPs)], as well as the use of smaller amplicons (< 300 bp). This study evaluated in 24 experimental runs the Precision ID GlobalFiler™ NGS STR panel v2 from ThermoFisher, which targets 31 autosomal STRs, amelogenin, and three Y-markers (one STR, SRY, and Yindel). Single-source samples were used in 18 experimental runs, for systematic evaluation. These included assessing library preparation benchmark conditions, limited DNA input, as well as testing repeatability, number of samples per run, and degraded DNA samples. Full profiles were consistently obtained from as little as 50 pg DNA input. Using the optional recovery PCR method improved outcomes for samples with low DNA input. Full profiles were also obtained from severely degraded DNA samples with degradation indices (DI) of > 60. In addition, six experimental runs were performed testing various two-person mixtures with mixture ratios ranging from 1:20 to 20:1. Major and minor contributors were distinguishable by their read counts (coverage), because less DNA input yielded lower read counts, analogous to the traditional CE technology, where less DNA produces lower peak heights. Mixture ratios of approximately 1:1 were indistinguishable, while a greater imbalance, i.e., higher mixture ratios, made the mixture more distinguishable between major and minor contributors. Based on this information, the highest success rate of correctly deconvoluted four-allelic loci was from mixtures with 1:3 ratios. At higher mixture ratios, the drop-out rate of the minor contributor increased, reducing the number of four-allelic loci.

Evaluation of ArmedXpert software tools, MixtureAce and Mixture Interpretation, to analyze MPS-STR data.

Massively parallel sequencing (MPS) technologies have revolutionized studies of genomic variations and transformed DNA analysis in multiple fields. Assays based on MPS must be capable of discriminating variations introduced by the method, i.e. artifacts from true polymorphisms. In PCR-MPS methods targeting microsatellite markers, artifacts can arise from PCR mis-incorporation, PCR strand slippage (stutter), and sequencing error. Reliable detection of artifacts in mixed DNA samples is a significant challenge that must be addressed in forensic DNA analysis. The ArmedXpert (NicheVision) software tools, MixtureAce™ and Mixture Interpretation, can analyze MPS data by categorizing sequence reads in alleles, stutter, and non-stutter artifacts and analyzing autosomal STR loci of mixed samples. In this study, we evaluated the ArmedXpert tools for the analysis of STR profiles of single-sourced and mixed samples generated by the ForenSeq™ DNA Signature Prep kit (Verogen). Data from eight experimental runs (240 samples) were analyzed: one benchmark run, two runs testing sensitivity with down to 50 pg DNA input, one run testing artificially degraded samples and DNA derived from bones, blood cards and teeth, as well as four runs with mixed DNA samples of varying ratios, sex, and different number of contributors (two to six). The MixtureAce stutter thresholds were initially set following the recommendations from Verogen, plus a non-stutter artifact threshold was set at 5% of allele read counts. A benchmark run, of 30 samples, plus two controls, containing 2310 total alleles, revealed over 5000 artifacts, above an analytical threshold of 10. A total of 4869 artifacts were correctly classified, while 435 were mis-classified as alleles due to exceedance of initial threshold settings. False positives must be resolved by an analyst, which can be time consuming. Stutter thresholds were adjusted based on the benchmark data and the samples were re-tested, resulting in only 57 false positive allele calls. The revised settings were then used in the analysis of the remaining seven experimental runs. Results show that MixtureAce can accurately classify artifacts and alleles when laboratory-specific threshold settings are used. The Mixture Interpretation tool was applied on two- and three-person mixtures. This tool utilized the analyzed data from MixtureAce to calculate, based on the number of alleles at a locus and their read counts, possible deconvolution outcomes with their respective ratios.

Micromanipulation of single cells and fingerprints for forensic identification.

Crime scene samples often include biological stains, handled items, or worn clothes and may contain cells from various donors. Applying routine sample collection methods by using a portion of a biological stain or swabbing the entire suspected touched area of the evidence followed by DNA extraction often leads to DNA mixtures. Some mixtures can be addressed with sophisticated interpretation protocols and probabilistic genotyping software resulting in DNA profiles of their contributors. However, many samples remain unresolved, providing no investigative information. Samples with many contributors are often the most challenging samples in forensic biology. Examples include gang rape situations or where the perpetrator's DNA is present in traces among the overwhelming amounts of the victim's DNA. If this is the only available evidence in a case, it is of paramount importance to generate usable information. An alternative approach, to address biological mixtures, could be the collection of individual cells directly from the evidence and testing them separately. This method could prevent cells from being inadvertently blended during the extraction process, thus resulting in DNA mixtures. In this study, multiple tools coupled with adhesive microcarriers to collect single cells were evaluated. These were tested on epithelial (buccal) and sperm cells, as well as on touched items. Single cells were successfully collected but fingerprints were swabbed in their entirety to account for the extracellular DNA of these samples and the poor DNA quality of shed skin flakes. Furthermore, micromanipulation devices, such as the P.A.L.M.® and the Axio Zoom.V16 operated manually or with a robotic arm aureka®, were compared for their effectiveness in collecting cells. The P.A.L.M.® was suitable for single cell isolation when smeared on membrane slides. Manual or robotic manipulations, by utilizing the Axio Zoom.V16, have wider applications as they can be used to isolate cells from various substrates such as glass or membrane slides, tapes, or directly from the evidence. Manipulations using the Axio Zoom.V16, either with the robotic arm aureka® or manually, generated similar outcomes which were significantly better than the outcomes by using the P.A.L.M.®. Robotic manipulations using the aureka® produced more consistent results, but operating the aureka® required training and often needed re-calibrations. This made the process of cell manipulations slower than when manually operated. Our preferred method was the manual manipulations as it was fast, cost effective, required little training, but relied on a steady hand of the technician.

Analyzing degraded DNA and challenging samples using the ForenSeq™ DNA Signature Prep kit.

Typing short tandem repeats (STRs) is the basis for human identification in current forensic testing. The standard method uses capillary electrophoresis (CE) to separate amplicons by length and fluorescent labeling. In recent years new methods, including massively parallel sequencing (MPS), have been developed which increased the discriminative power of STRs through sequencing. MPS also offers the opportunity to test more genetic markers in a run than is possible with standard CE technology. Verogen's ForenSeq™ DNA Signature Prep kit includes over 150 genetic markers [STRs and single nucleotide polymorphisms (SNPs)]. Further, MPS separation depends on sequences rather than lengths; therefore, amplicons can be small or even of the same lengths. These improvements are advantageous when testing challenging forensic samples that could be severely degraded. This study tested the ForenSeq™ DNA Signature Prep kit in repeated experimental runs on series of degraded DNA samples, ranging from mild to severe degradation, as well as 24 mock case-type samples, derived from bones, blood cards, and teeth. Despite passing the quality metrics, positive controls (2800 M) showed drop-outs at some loci, mostly SNPs. Sequencing DNA samples repeatedly in two experimental runs as well as sequencing one pooled library in triplicate led to the assumption that spurious alleles of the Y-STRs in this study were not a result of sequencing artifacts but could be due to sequence structures (e.g. duplications, palindromes) of the Y-chromosome and/or might be accumulated during library preparation. Two sets of serially degraded DNA samples revealed that dropped-out loci were primarily loci with long amplicons as well as low read numbers (coverage), e.g. PentaE, DXS8378, and rs1736442. STRs started to drop out at degradation indices (DIs) > 4. However, severely degraded DNA (DI: 44) still resulted in 90% of the 20 CODIS loci, while only 35% were obtained using Promega's PowerPlex® Fusion kit, a current standard CE kit. Mock case-type samples confirmed these results. ForenSeq™ DNA Signature Prep kit demonstrated that it can be successfully used on degraded DNA samples. This study may be helpful for other laboratories assessing and validating MPS technologies.

Evaluation of ForenSeq™ Signature Prep Kit B on predicting eye and hair coloration as well as biogeographical ancestry by using Universal Analysis Software (UAS) and available web-tools.

This study examined 266 individuals from various populations including African American, East Asian, South Asian, European, and mixed populations to evaluate the ForenSeq™ Signature Prep Kit Primer Mix B. Focus was placed on phenotypic and biogeographical ancestry predictions by Illumina's Universal Analysis Software (UAS). These outcomes were compared to those obtained through web-tools developed at the Erasmus Medical Center (EMC) and available from the Forensic Resource/Reference on Genetics-knowledge base (FROG-kb), as well as to eye color predictions by the 8-plex system. Due to drop-outs, predictions for eye and hair color by UAS failed for various samples in each run. By including reads below thresholds, predictions could be obtained for all samples through the web-tools. Eye and hair color predictions for African Americans, East Asians, and South Asians showed no errors. Difficulties however, were noted in intermediate (neither blue nor brown) eye color predictions. These were mitigated by the 8-plex system through exclusion of one eye color (e.g. "not brown"). Additionally, notable discrepancies were observed in hair color predictions, where some black/dark-brown haired individuals were predicted to have blond hair. Overall, ancestry predictions were more accurate by FROG-kb compared to UAS, which did not predict South Asian ancestry, particularly Indian individuals.

Qualitative and quantitative assessment of Illumina's forensic STR and SNP kits on MiSeq FGx™.

Massively parallel sequencing (MPS) is a powerful tool transforming DNA analysis in multiple fields ranging from medicine, to environmental science, to evolutionary biology. In forensic applications, MPS offers the ability to significantly increase the discriminatory power of human identification as well as aid in mixture deconvolution. However, before the benefits of any new technology can be employed, a thorough evaluation of its quality, consistency, sensitivity, and specificity must be rigorously evaluated in order to gain a detailed understanding of the technique including sources of error, error rates, and other restrictions/limitations. This extensive study assessed the performance of Illumina's MiSeq FGx MPS system and ForenSeq™ kit in nine experimental runs including 314 reaction samples. In-depth data analysis evaluated the consequences of different assay conditions on test results. Variables included: sample numbers per run, targets per run, DNA input per sample, and replications. Results are presented as heat maps revealing patterns for each locus. Data analysis focused on read numbers (allele coverage), drop-outs, drop-ins, and sequence analysis. The study revealed that loci with high read numbers performed better and resulted in fewer drop-outs and well balanced heterozygous alleles. Several loci were prone to drop-outs which led to falsely typed homozygotes and therefore to genotype errors. Sequence analysis of allele drop-in typically revealed a single nucleotide change (deletion, insertion, or substitution). Analyses of sequences, no template controls, and spurious alleles suggest no contamination during library preparation, pooling, and sequencing, but indicate that sequencing or PCR errors may have occurred due to DNA polymerase infidelities. Finally, we found utilizing Illumina's FGx System at recommended conditions does not guarantee 100% outcomes for all samples tested, including the positive control, and required manual editing due to low read numbers and/or allele drop-in. These findings are important for progressing towards implementation of MPS in forensic DNA testing.

Analysis of fingerprint samples, testing various conditions, for forensic DNA identification.

Fingerprints can be of tremendous value for forensic biology, since they can be collected from a wide variety of evident types, such as handles of weapons, tools collected in criminal cases, and objects with no apparent staining. DNA obtained from fingerprints varies greatly in quality and quantity, which ultimately affects the quality of the resulting STR profiles. Additional difficulties can arise when fingerprint samples show mixed STR profiles due to the handling of multiple persons. After applying a tested protocol for sample collection (swabbing with 5% Triton X-100), DNA extraction (using an enzyme that works at elevated temperatures), and PCR amplification (AmpFlSTR® Identifiler® using 31cycles) extensive analysis was performed to better understand the challenges inherent to fingerprint samples, with the ultimate goal of developing valuable profiles (≥50% complete). The impact of time on deposited fingerprints was investigated, revealing that while the quality of profiles deteriorated, full STR profiles could still be obtained from samples after 40days of storage at room temperature. By comparing the STR profiles from fingerprints of the dominant versus the non-dominant hand, we found a slightly better quality from the non-dominant hand, which was not always significant. Substrates seem to have greater effects on fingerprints. Tests on glass, plastic, paper and metal (US Quarter dollar, made of Cu and Ni), common substrates in offices and homes, showed best results for glass, followed by plastic and paper, while almost no profiles were obtained from a Quarter dollar. Important for forensic casework, we also assessed three-person mixtures of touched fingerprint samples. Unlike routinely used approaches for sampling evidence, the surface of an object (bottle) was sectioned into six equal parts and separate samples were taken from each section. The samples were processed separately for DNA extraction and STR amplification. The results included a few single source profiles and distinguishable two person mixtures. On average, this approach led to two profiles ≥50% complete per touched object. Some STR profiles were obtained more than once thereby increasing the confidence.

Systematic assessment of the performance of illumina's MiSeq FGx™ forensic genomics system.

This study assesses the performance of Illumina's MiSeq FGx System for forensic genomics by systematically analyzing single source samples, evaluating concordance, sensitivity and repeatability, as well as describing the quality of the reported outcomes. DNA from 16 individuals (9 males/7 females) in nine separate runs showed consistent STR profiles at DNA input ≥400 pg, and two full profiles were obtained with 50 pg DNA input. However, this study revealed that the outcome of a single sample does not merely depend on its DNA input but is also influenced by the total amount of DNA loaded onto the flow cell from all samples. Stutter and sequence or amplification errors can make the identification of true alleles difficult, particularly for heterozygous loci that show allele imbalance. Sequencing of 16 individuals' STRs revealed genetic variations at 14 loci at frequencies suggesting improvement of mixture deconvolution. The STR loci D1S1656 and DXS10103 were most susceptible to drop outs, and D22S1045 and DYS385a-b showed heterozygote imbalance.  Most stutters were typed at TH01 and DYS385a-b, while amplification or sequencing errors were observed mostly at D7S820 and D19S433. Overall, Illumina's MiSeq FGx System produced reliable and repeatable results.  aSTRs showed fewer drop outs than the Y- and X-STRs.

Assay Development and Validation of an 8-SNP Multiplex Test to Predict Eye and Skin Coloration.

Identifying human remains is one of the many responsibilities of forensic scientists. An eye- and skin-color predictor translates genotypic information into phenotypic description. Eight single nucleotide polymorphisms (SNPs) are utilized for this predictor, five for eye, and six for skin coloration. Here, we describe the development and validation of an 8-SNP multiplex assay that consists of a multiplex PCR, followed by a multiplexed single-base primer extension reaction generating fluorescently labeled oligonucleotides of distinct length that are detected by multicolor capillary electrophoresis. Validation of this assay included tests for reproducibility, reliability, sensitivity, species specificity, its performance on degraded DNA, and on forensic samples. It can be concluded that the 8-SNP multiplex assay is robust and can be used on challenging samples, including bones, to reliably determine the genotypes to predict eye and skin color of individuals. This information can assist in the identification of human remains and missing persons.

Qualitative and quantitative assessment of single fingerprints in forensic DNA analysis.

Fingerprints and touched items are important sources of DNA for STR profiling, since this evidence can be recovered in a wide variety of criminal offenses. However, there are some fundamental difficulties in working with these samples, including variability in quantity and quality of extracted DNA. In this study, we collected and analyzed over 700 fingerprints. We compared a commercially available extraction protocol (Zygem) to two methods developed in our laboratory, a simple one-tube protocol and a high sensitivity protocol (HighSens) that includes additional steps to concentrate and purify the DNA. The amplification protocols tested were AmpFLSTR® Identifiler® using either 28 or 31 amplification cycles, and Identifiler® Plus using 32 amplification cycles. We found that the HighSens and Zygem extraction methods were significantly better in their DNA yields than the one-tube method. Identifiler® Plus increased the quality of the STR profiles for the one-tube extraction significantly. However, this effect could not be verified for the other extraction methods. Furthermore, microscopic analysis of single fingerprints revealed that some individuals tended to shed more material than others onto glass slides. However, a dense deposition of skin flakes did not strongly correlate with a high quality STR profile.

Deletion mapping in the Enhancer of split complex.

The Enhancer of split complex [E(spl)-C] comprises twelve genes of different classes. Seven genes encode proteins of with a basic-helix-loop-helix-orange (bHLH-O) domain that function as transcriptional repressors and serve as effectors of the Notch signalling pathway. They have been named E(spl)m8-, m7-, m5-, m3-, mß-, mγ- and mδ-HLH. Four genes, E(spl)m6-, m4-, m2- and mα-BFM are intermingled and encode Notch repressor proteins of the Bearded-family (BFM). The complex is split by a single gene of unrelated function, encoding a Kazal-type protease inhibitor (Kaz-m1). All members within a family, bHLH-O or BFM, are very similar in structure and in function. In an attempt to generate specific mutants, we have mobilised P-element constructs residing next to E(spl)m7-HLH and E(spl)mγ-HLH, respectively. The resulting deletions were mapped molecularly and by cytology. Two small deletions affected only E(spl)m7-HLH and E(spl)mδ. The deficient flies were viable without apparent phenotype. Larger deletions, generated also by X-ray mutagenesis, uncover most of the E(spl)-C. The phenotypes of homozygous deficient embryos were analysed to characterize the respective loss of Notch signalling activity.

Improved eye- and skin-color prediction based on 8 SNPs.

AIM: To improve the 7-plex system to predict eye and skin color by increasing precision and detailed phenotypic descriptions. METHODS: Analysis of an eighth single nucleotide polymorphism (SNP), rs12896399 (SLC24A4), showed a statistically significant association with human eye color (P=0.007) but a rather poor strength of agreement (κ=0.063). This SNP was added to the 7-plex system (rs12913832 at HERC2, rs1545397 at OCA2, rs16891982 at SLC45A2, rs1426654 at SLC24A5, rs885479 at MC1R, rs6119471 at ASIP, and rs12203592 at IRF4). Further, the instruction guidelines on the interpretation of genotypes were changed to create a new 8-plex system. This was based on the analysis of an 803-sample training set of various populations. The newly developed 8-plex system can predict the eye colors brown, green, and blue, and skin colors light, not dark, and not light. It is superior to the 7-plex system with its additional ability to predict blue eye and light skin color. RESULTS: The 8-plex system was tested on an additional 212 samples, the test set. Analysis showed that the number of positive descriptions for eye colors as being brown, green, or blue increased significantly (P=6.98e-15, z-score: -7.786). The error rate for eye-color prediction was low, at approximately 5%, while the skin color prediction showed no error in the test set (1% in training set). CONCLUSIONS: We can conclude that the new 8-plex system for the prediction of eye and skin color substantially enhances its former version.

Verification of eye and skin color predictors in various populations.

Validation of testing methods is an essential feature in all scientific endeavors, but it is particularly important in forensics. Due to the sensitive nature of these investigations and the limited sample size it is crucial to validate all employed procedures. This includes novel forensic phenotypic DNA tests, to learn more of their capabilities and limitations before incorporating them as routine methods. Ideally, validations are performed on large sample sets that mimic real cases. Recently, three phenotypic predictors, two for eye colors and one for skin color have been published (Spichenok et al., 2011; Walsh et al., 2011). These predictors are well-defined by a selection of single nucleotide polymorphisms (SNPs) and unambiguous instructions on how to interpret the genotypes. These standardized approaches have the advantages that they can be applied in diverse laboratories leading to the same outcome and offer the opportunity for validation. For these tests to be used on the characterization of human remains, they should be validated on various populations to perform reliably without prior knowledge of ethnic origin. Here, in this study, these eye and skin color predictors were validated on new sample sets and it could be confirmed that they can be applied in various populations, including African-American, South Asian (dark), East Asian (light), European, and mixed populations. The outputs were either predictive or inconclusive. Predictions were then compared against the actual eye and skin colors of the tested individuals. The error-rates varied; they were low for the predictors that describe the eye and skin color exclusively (non-brown or non-blue and non-white or non-dark, respectively) and higher for the predictor that describes individual eye colors (blue, brown, and intermediate/green), because of uncertainties with the green eye color prediction. Our investigation deepens the insight for these predictors and adds new information.

Prediction of eye and skin color in diverse populations using seven SNPs.

An essential component in identifying human remains is the documentation of the decedent's visible characteristics, such as eye, hair and skin color. However, if a decedent is decomposed or only skeletal remains are found, this critical, visibly identifying information is lost. It would be beneficial to use genetic information to reveal these visible characteristics. In this study, seven single nucleotide polymorphisms (SNPs), located in and nearby genes known for their important role in pigmentation, were validated on 554 samples, donated from non-related individuals of various populations. Six SNPs were used in predicting the eye color of an individual, and all seven were used to describe the skin coloration. The outcome revealed that these markers can be applied to all populations with very low error rates. However, the call-rate to determine the skin coloration varied between populations, demonstrating its complexity. Overall, these results prove the importance of these seven SNPs for potential forensic tests.

De-regulation of common housekeeping genes in hepatocellular carcinoma.

BACKGROUND: Tumorigenesis is associated with changes in gene expression and involves many pathways. Dysregulated genes include "housekeeping" genes that are often used for normalization for quantitative real-time RT-PCR (qPCR), which may lead to unreliable results. This study assessed eight stages of hepatitis C virus (HCV) induced hepatocellular carcinoma (HCC) to search for appropriate genes for normalization. RESULTS: Gene expression profiles using microarrays revealed differential expression of most "housekeeping" genes during the course of HCV-HCC, including glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and beta-actin (ACTB), genes frequently used for normalization. QPCR reactions confirmed the regulation of these genes. Using them for normalization had strong effects on the extent of differential expressed genes, leading to misinterpretation of the results. CONCLUSION: As shown here in the case of HCV-induced HCC, the most constantly expressed gene is the arginine/serine-rich splicing factor 4 (SFRS4). The utilization of at least two genes for normalization is robust and advantageous, because they can compensate for slight differences of their expression when not co-regulated. The combination of ribosomal protein large 41 (RPL41) and SFRS4 used for normalization led to very similar results as SFRS4 alone and is a very good choice for reference in this disease as shown on four differentially expressed genes.

Genome-wide molecular profiles of HCV-induced dysplasia and hepatocellular carcinoma.

UNLABELLED: Although HCC is the third-leading cause of cancer-related deaths worldwide, there is only an elemental understanding of its molecular pathogenesis. In western countries, HCV infection is the main etiology underlying this cancer's accelerating incidence. To characterize the molecular events of the hepatocarcinogenic process, and to identify new biomarkers for early HCC, the gene expression profiles of 75 tissue samples were analyzed representing the stepwise carcinogenic process from preneoplastic lesions (cirrhosis and dysplasia) to HCC, including 4 neoplastic stages (very early HCC to metastatic tumors) from patients with HCV infection. We identified gene signatures that accurately reflect the pathological progression of disease at each stage. Eight genes distinguish between control and cirrhosis, 24 between cirrhosis and dysplasia, 93 between dysplasia and early HCC, and 9 between early and advanced HCC. Using quantitative real-time reverse-transcription PCR, we validated several novel molecular tissue markers for early HCC diagnosis, specifically induction of abnormal spindle-like, microcephaly-associated protein, hyaluronan-mediated motility receptor, primase 1, erythropoietin, and neuregulin 1. In addition, pathway analysis revealed dysregulation of the Notch and Toll-like receptor pathways in cirrhosis, followed by deregulation of several components of the Jak/STAT pathway in early carcinogenesis, then upregulation of genes involved in DNA replication and repair and cell cycle in late cancerous stages. CONCLUSION: These findings provide a comprehensive molecular portrait of genomic changes in progressive HCV-related HCC.

A molecular signature to discriminate dysplastic nodules from early hepatocellular carcinoma in HCV cirrhosis.

BACKGROUND & AIMS: Small liver nodules approximately 2 cm are difficult to characterize by radiologic or pathologic examination. Our aim was to identify a molecular signature to diagnose early hepatocellular carcinoma (HCC). METHODS: The transcriptional profiles of 55 candidate genes were assessed by quantitative real-time reverse-transcription polymerase chain reaction (RT-PCR) in 17 dysplastic nodules (diameter, 10 mm) and 20 early HCC (diameter, 18 mm) from HCV cirrhotic patients undergoing resection/transplantation and 10 nontumoral cirrhotic tissues and 10 normal liver tissues. Candidate genes were confirmed by quantitative RT-PCR in 20 advanced HCCs and by immunohistochemistry in 75 samples and validated in an independent set of 29 samples (dysplastic nodules [10] and small HCC [19; diameter, 20 mm]). RESULTS: Twelve genes were significantly, differentially expressed in early HCCs compared with dysplastic nodules (>2-fold change; area under the receiver operating characteristic curve > or =0.8): this included TERT, GPC3, gankyrin, survivin, TOP2A, LYVE1, E-cadherin, IGFBP3, PDGFRA, TGFA, cyclin D1, and HGF. Logistic regression analysis identified a 3-gene set including GPC3 (18-fold increase in HCC, P = .01), LYVE1 (12-fold decrease in HCC, P = .0001), and survivin (2.2-fold increase in HCC, P = .02), which had a discriminative accuracy of 94%. The validity of the gene signature was confirmed in a prospective testing set. GPC3 immunostaining was positive in all HCCs and negative in dysplastic nodules (22/22 vs 0/14, respectively, P < .001). Nuclear staining for survivin was positive in 12 of 13 advanced HCC cases and in 1 of 9 early tumors. CONCLUSIONS: Molecular data based on gene transcriptional profiles of a 3-gene set allow a reliable diagnosis of early HCC. Immunostaining of GPC3 confirms the diagnosis of HCC.

Acute induction of gene expression in brain and liver by insulin-induced hypoglycemia.

The robust neuroendocrine counterregulatory responses induced by hypoglycemia protect the brain by restoring plasma glucose, but little is known about molecular responses to hypoglycemia that may also be neuroprotective. To clarify these mechanisms, we examined gene expression in hypothalamus, cortex, and liver 3 h after induction of mild hypoglycemia by a single injection of insulin, using cDNA microarray analysis and quantitative real-time PCR. Real-time PCR corroborated the induction of six genes (angiotensinogen, GLUT-1, inhibitor of kappaB, inhibitor of DNA binding 1 [ID-1], Ubp41, and mitogen-activated protein kinase phosphatase-1 [MKP-1]) by insulin-induced hypoglycemia in the hypothalamus: five of these six genes in cortex and three (GLUT-1, angiotensinogen, and MKP-1) in liver. The induction was due to hypoglycemia and not hyperinsulinemia, since fasting (characterized by low insulin and glucose) also induced these genes. Four of these genes (angiotensinogen, GLUT-1, ID-1, and MKP-1) have been implicated in enhancement of glucose availability, which could plausibly serve a neuroprotective role during acute hypoglycemia but, if persistent, could also cause glucose-sensing mechanisms to overestimate plasma glucose levels, potentially causing hypoglycemia-induced counterregulatory failure. Although using cDNA microarrays with more genes, or microdissection, would presumably reveal further responses to hypoglycemia, these hypoglycemia-induced genes represent useful markers to assess molecular mechanisms mediating cellular responses to hypoglycemia.

Mining microarrays for metabolic meaning: nutritional regulation of hypothalamic gene expression.

DNA microarray analysis has been used to investigate relative changes in the level of gene expression in the CNS, including changes that are associated with disease, injury, psychiatric disorders, drug exposure or withdrawal, and memory formation. We have used oligonucleotide microarrays to identify hypothalamic genes that respond to nutritional manipulation. In addition to commonly used microarray analysis based on criteria such as fold-regulation, we have also found that simply carrying out multiple t tests then sorting by P value constitutes a highly reliable method to detect true regulation, as assessed by real-time polymerase chain reaction (PCR), even for relatively low abundance genes or relatively low magnitude of regulation. Such analyses directly suggested novel mechanisms that mediate effects of nutritional state on neuroendocrine function and are being used to identify regulated gene products that may elucidate the metabolic pathology of obese ob/ob, lean Vgf-/Vgf-, and other models with profound metabolic impairments.